Originally posted by: JG. KIM

Hi, everyone.

I'm new at CP. So, I'm developing a model to schedule Energy charging & discharging on ESS battery.

So, I need your help. please help me :)

 I have 5 batteries and based on time charging, discharging costs.

these batteries have the initial capacity. this capacity will be changed based on time because of acting charge or discharge.

So, I want to make 3 decision variables.

1. charging, discharging signal. (sig)

2. charging, discharging power (Cpower, Dpower)

but, when I make code like this : I can see 'no solution'

using CP;


int numEVs = ...;
range EVs = 0..numEVs-1;
int time = ...;
range times = 0..time-1;
int k[times,EVs]=...;

int cost[times] = [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24];
float min_soc[EVs] = [0.4,0.4,0.4,0.4,0.4];
float max_soc[EVs] = [0.9,0.9,0.9,0.9,0.9];
float Soc[EVs] = [0.4, 0.5, 0.6, 0.7, 0.8];

tuple EVs2 {
  key int id;
  int Cpower[times];
  int Dpower[times];
}

//float delSm[EVs] = Soc[EVs] - min_soc[EVs]; 
//float delSp[EVs] = Soc[EVs] - max_soc[EVs];
dvar interval t[i in times] optional size 1;
dvar int Pcmax[times, EVs]; // why can't use float.
dvar int Pdmax[times, EVs];

dexpr float Cost = sum(t, j in EVs) (k[t,j]*cost[t]*Pcmax[t,j] - (1-k[t,j])*cost[t]*Pdmax[t,j]);

minimize Cost; // minimize charging/discharging price

subject to {
  forall(t, j in EVs)
  k[t,j]*Pcmax[t,j] - (1-k[t,j])*Pdmax[t,j] >= Soc[j]-min_soc[j] && k[t,j]*Pcmax[t,j] - (1-k[t,j])*Pdmax[t,j] <= Soc[j]-max_soc[j];  
// each EV's battery state of charge must less or bigger than limits. 
  forall(t, j in EVs)
    k[t,j] == 1 || k[t,j] ==0;
  forall(t, j in EVs)
    {
    Pdmax[t][j] >=0;
    Pdmax[t][j] <=10;
    Pcmax[t][j] >=0;
    Pcmax[t][j] <=8;
        }
}

And I want to add constraints about changing capacity. how can I model this?

Thank you for help.

Originally posted by: ChrisBr

Hello,

You might try the Conflict Refiner ; it will help you to find which constraints lead to this result with no solution.
The simplest way would be:

oplrun -conflict model.mod data.dat

See documentation of oplrun Command Line Interface.

I hope this helps,

Chris.

Originally posted by: JG. KIM

Thank you for answer. :)

but I still have problem..

Can we make 1-day battery charging/discharging scheduling with CP?

Originally posted by: Petr Vilím

Hello,

could you give us also your data file? Without it I cannot test your model.

I noticed that there are interval variables "t" in the model. However they are not used (there are more identifiers "t" in the model used as indexes in the forall and they hide "t" from outer scopes). So I recommend to delete those variables (but it will not solve the problem probably).

I'm also wondering about the following constraint:

  k[t,j]*Pcmax[t,j] - (1-k[t,j])*Pdmax[t,j] >= Soc[j]-min_soc[j] && k[t,j]*Pcmax[t,j] - (1-k[t,j])*Pdmax[t,j] <= Soc[j]-max_soc[j]; 

It is in the form A >= B <= C. I'm wondering whether it isn't a typo, maybe it was intended as A <= B <= C?

If it is not a typo then maybe you can rewrite it to B <= minl(A, C).

In the model you have a question why you cannot use floating-point decision variables. It is simply not supported in CP. You can scale an integer variable instead. For example, if precision of 2 decimal places is enough then you can use integer variable x and use x/100 everywhere.

Best regards, Petr

Originally posted by: Petr Vilím

I was able to run your model using the data file from your second question. And here are two constraints that together make the problem infeasible:

"Pcmax[3,2]" + -("Pdmax[3,2]" * (1 + -1)) <= 0.6 + -0.9;
"Pcmax[3,2]" >= 0;

Best regards, Petr